It is axiomatic in the investment world that as an asset class becomes more popular, it suffers from both falling expected returns and rising correlations. In other words, good trades get crowded and their advantages tend to disappear. This crowding happens because success begets copycats as investors chase returns. Mean reversion only tends to make matters worse. In effect, it results in “investing while looking in the rearview mirror” or, as per the title of William Bernstein’s fine new book, Skating Where the Puck Was.
The evidence suggests that this overcrowding is precisely what has been happening with respect to the Yale Model. Despite Yale’s annual net average investment return of 13.7 percent over the past 20 years, returns haven’t been nearly as spectacular of late and, significantly, the great bulk of Yale’s imitators have not done nearly as well as Yale. Those facts alone suggest overcrowding.
The Yale Model became the focus of much attention on account of its performance during the tech bubble. From July 2000 through June 2003, while the S&P 500 fell 33 percent, Yale’s endowment actually gained 20 percent. Yale’s portfolio continued to perform well until 2008 (returns of 19.4 percent in fiscal 2004, 22.3 percent in fiscal 2005, 22.9 percent in fiscal 2006, 28.0 percent in fiscal 2007 and 4.5 percent in fiscal 2008 – through June 30, 2008).
However, the poor performance of the fund and others like it during the financial crisis beginning in 2008 precipitated some intense questioning of the Yale Model. Yale lost 24.6 percent in its fiscal year 2009 (compared to an S&P 500 loss of roughly 26 percent over that time), considerably worse than the average endowment, which lost 18.6 percent over the same period. As a result, a report by the Tellus Institute and the Center for Social Philanthropy boldly called for “a transformation of the Endowment Model of Investing” on account of what it perceived to be excessive risk-taking. Moody’s issued a report with similarly solemn tones, particularly criticizing the Model’s liquidity management. The Yale Model received its share of academic criticism too.
Obviously, one difficult patch (even a really difficult patch) is hardly sufficient evidence to discredit Swensen’s approach in toto. The idea that a broadly diversified portfolio is a low-risk portfolio is common but wrong. A well-diversified portfolio has higher average expected return for a lower average risk level, but that is not to say that it is somehow sheltered from significant drawdowns. Swensen’s own research (outlined in the Endowment’s 2010 Report) discloses the extremity of the risks associated with the Yale Model. Swensen concedes that the “average endowment” runs a 28 percent likelihood of losing fully half of its assets (in real terms) over the next 50 years based on its asset allocations and a 35 percent risk of a “spending disruption” over the next five years. Clearly, the Yale Model postulates a very high risk portfolio.
However, Yale’s performance since the crisis has not regained its prior heights. Indeed, it has been decidedly lackluster, suggesting more than a temporary down cycle (mean reversion) is at work. In fiscal 2010, Yale achieved 8.9 percent return while the S&P 500 gained 12 percent. In fiscal 2011, Yale earned 21.9 percent, roughly even with the S&P 500. And in, fiscal 2012 (ending June 30, 2012), the most recent period for which data is available, Yale earned 4.7 percent, less than the S&P 500’s 5.5 percent over that time (the big equity market rally kicked in during the second half of 2012, after the reporting period).
That said, Yale’s longer-term results remain in the top-tier of institutional investors. Yale’s endowment earned 10.6 percent per annum over the 10 years ending June 30, 2012, well lower than 20-year return numbers but well surpassing results for domestic stocks, which returned 6.2 percent annually, domestic bonds, which returned 5.6 percent annually and college and university endowments generally, which on average returned 6.8 percent annually.
Even so, a closer look at the data, and particularly at Yale’s imitators, suggests that the Yale approach may have passed its peak. As per the Yale Model, most endowments, many institutions and a sizable number of individual investors have sought to add alternative investments to their portfolios. The average endowment holds roughly 15 percent of its assets in U.S. stocks as compared with 54 percent in alternatives. As a diversifier, non-correlated assets cannot be expected to perform when traditional assets do. However, many of these investments are sought out for their return characteristics rather than as diversifiers. Moreover, many of these alternatives are positioned as absolute return vehicles, nimble enough to provide positive returns in any environment. The data suggests that neither approach is all that successful.
Indeed, data compiled by the National Association of College and University Business Officers for the most recent fiscal year show that large, medium and small endowments have all underperformed in recent years. Nacubo and Commonfund gathered data from 831 U.S. colleges and universities managing over $400 billion in assets. The average return for U.S. college endowments was minus-0.3 percent in the 12 months ended in June 2012 (the most recent period for which data is available). Longer-term returns are not a lot better. College and university endowments returned an average of only 1.1 percent annually over the past five fiscal years and 6.2 percent over the past decade, net of fees.
A simple 60/40 portfolio invested 60 percent in an S&P 500 index fund and 40 percent in a fund tracking the Barclays Aggregate bond index would have gained 12.6 percent annually over the last three years and 2.8 percent over the last five years, as compared with 10.2 percent and 1.1 percent, respectively, for the average endowment. Yale did somewhat better than the average endowment with three-year returns averaging 11.83 percent and five-year returns averaging 3.08 percent.
Significantly, the largest endowments, those with more than $1 billion in assets (not to mention the most resources generally), posted the best returns, gaining 0.8 percent in fiscal 2012 and an average of 7.6 percent annually over the past 10 years. Endowments with $51 million to $100 million in assets lost 1 percent last year and returned an average of 5.7 percent annually for the past 10 years.
Even more significantly, more than 90 percent of endowments with less than $1 billion in assets underperformed in every time period since such records have been kept. Since multiple reports suggest that current year returns have been even more problematic, this trend is likely to get worse before it gets better, largely due to the rally in stocks commencing during the second half of calendar 2012. To the extent that the Yale Model remains strong, its value is heavily concentrated in the largest endowments, suggesting that smaller institutions and individual investors are particularly at risk if they try to emulate Yale. As my friend Tadas Viskanta wrote recently, “institutions have capabilities, access and expertise that individuals simply don’t have.”
Here are some actual numbers, with a bit of tweaking to the 60/40 portfolios, demonstrating the outperformance of 60/40 portfolios versus the average college endowment over three, five and ten years as well as the better drawdown protection they provided during the financial crisis.
60/40 Index Portfolio = 42% Russell 3000 Index, 18% MSCI All Country World ex. US Index, 40% Barclays Aggregate Bond Index, rebalanced annually
60/40 Asset Class Portfolio = 9% S&P 500 Index, 12% DFA US large value index, 21% DFA US targeted value index, 6% DFA international large value index, 6% DFA international small value index, 6% DFA emerging market value index, 40% Ibbotson 5YR T-Note index, rebalanced annually
This problem is not limited to endowments, consistent with the overcrowding hypothesis. State pension fund trustees are just as bad as university endowments when implementing the Yale model in that pension returns also fall short of a simple index fund portfolio.
Thus the return numbers alone suggest a very strong possibility that the Yale approach may well be overcrowded, with ongoing outperformance difficult to achieve. Further analysis supports that conclusion as well. Multiple studies (for example, here) together with Swensen’s own analysis have determined that Yale’s outperformance has largely been predicated upon the success of its private equity exposure together with a (related) illiquidity premium. Research indicates that heavy exposure to common equity risk factors and manager skill in private equity drove Yale’s returns. Although the liquidity risk premium is significant (about 3 percent annually – consistent with a growing body of research that supports the idea that illiquid assets provide the potential for higher return), private equity returns are pretty well explained by their exposure to the overall risk of the stock market, the risks of value stocks, the risks of investing in less liquid securities and leverage.
Moreover, the illiquidity comes at a cost. When Harvard tried to get out of illiquid private equity funds in the depths of the financial crisis, it found it would have to take a haircut of 40-60 percent to do so and alternative assets were frequently liquidated at less than 50 percent of the previous year-end net asset value. Simply put, institutions with insufficient liquidity in 2008 got crushed. As a consequence, Swensen took pains to point out that he “never took the position liquidity wasn’t important, just that it’s generally overvalued.”
The bottom line here is that Yale’s success has been driven largely by private equity – mostly venture capital – returns. Outside of private equity, the research suggests that Yale appeared to underperform appropriate risk-adjusted benchmarks.
Significantly, endowment funds as a whole dramatically outperformed during most of the 1990s largely due to the success of their private equity investments. On the other hand, for funds raised during 1999-2006, endowments underperformed.
The 1990s outperformance was largely driven by endowment investments in older, more established venture capital firms which were frequently difficult to get into and size-constrained but with great results. By 1999-2006, the performance advantages of these types of funds had largely dissipated. In other words, the trade had gotten increasingly crowded. For further support of this idea, Cambridge Associates estimates that 3 percent of venture capital firms generate 95 percent of the industry’s returns and adds that there is little change in the composition of those 3 percent of firms over time (more here).
The table below from Abbott Capital (a private equity fund of funds) via Sober Look illustrates this history. The top half is a table that shows returns for various private equity sectors. For example, the 2000 “large buyout” funds produced an 18.6% annualized average return over their lives (typically 7-9 years). The lower half shows the dollar amounts raised for the two major private equity groups (venture and buy-out) over the same time period.
The upshot here, emphasized by the red circles immediately above, is that by the time large amounts of capital poured into the asset class, the trade was already too crowded and the party was over. Not surprisingly, endowment returns across-the-board reflect this same trend, as shown above. It’s a consistent problem – chasing returns and buying assets that have already peaked.
It is also highly significant that hedge funds are responsible for a very large proportion of endowment investments in alternative strategies. Simon Lack, a hedge fund insider from his long tenure at JPMorgan Chase, raised a real ruckus last year with his book, The Hedge Fund Mirage, in which he argued that the hedge fund industry as a whole lost more money in 2008 than it had made in the previous 10 years. Even worse, “[i]f all the money that’s ever been invested in hedge funds had been put in Treasury bills instead, the results would have been twice as good,” especially since nearly all the hedge funds’ gains had gone to managers rather than clients.
Of course, hedge funds are susceptible to overcrowding too and hedge funds now manage $2.2 trillion in assets, up fourfold since 2000. Even so, Swensen argues that “[h]igh quality hedge-fund exposure that produces returns that are fundamentally independent of what’s going on in the markets can be a great addition to an institutional portfolio.” However, in 2012, the HFRX Global Hedge Fund Index, reflective of industry returns, gained just 3.5 percent and has returned just 1.7 percent over the past 10 years. Thus the S&P 500 has outperformed the HFRX for ten straight years, with the exception of 2008 when both fell sharply. More significantly, a standard 60/40 portfolio has delivered returns of more than 90 percent over the past decade, compared with a meager 17 percent after fees for hedge funds, as shown below.
Source: The Economist
“If you look at the data, hedge funds have underperformed a simple 60/40 stock/bond mix every year for the past 10 years,” Lack asserts. “They did well in the downturn of 2000-2. But that’s when assets under management were less than half what they are now. There’s no disputing that as assets have grown, performance has declined.” No matter how you slice it, hedge fund performance has been disappointing and money is now migrating elsewhere. Hedge fund investing looks very crowded indeed.
The available evidence provides a pretty good case for the idea that the Yale Model is past its peak due to overcrowding. A more traditional approach may simply make more sense, even if such an approach is decidedly less “sexy,” especially for those who don’t have Yale’s access and expertise.